Reliable Multimodal Trajectory Prediction via Error Aligned Uncertainty Optimization

TL;DR

This paper introduces a novel error aligned uncertainty optimization method for deep neural networks, enhancing the quality of uncertainty estimates in safety-critical tasks like vehicle trajectory prediction, especially under real-world distributional shifts.

Contribution

The paper proposes a trainable loss function for better uncertainty calibration in continuous prediction tasks, improving correlation with model error and robustness in real-world scenarios.

Findings

Improves average displacement error by up to 4.69%.

Enhances uncertainty correlation with model error by over 19%.

Effective on multiple datasets, including large-scale vehicle prediction.

Abstract

Reliable uncertainty quantification in deep neural networks is very crucial in safety-critical applications such as automated driving for trustworthy and informed decision-making. Assessing the quality of uncertainty estimates is challenging as ground truth for uncertainty estimates is not available. Ideally, in a well-calibrated model, uncertainty estimates should perfectly correlate with model error. We propose a novel error aligned uncertainty optimization method and introduce a trainable loss function to guide the models to yield good quality uncertainty estimates aligning with the model error. Our approach targets continuous structured prediction and regression tasks, and is evaluated on multiple datasets including a large-scale vehicle motion prediction task involving real-world distributional shifts. We demonstrate that our method improves average displacement error by 1.69% and 4.69%, and the uncertainty correlation with model error by 17.22% and 19.13% as quantified by Pearson correlation coefficient on two state-of-the-art baselines.